LUND UNIVERSITY

PO Box 117221 00 Lund+46 46-222 00 00

Economic stress in childhood and adulthood, and poor psychological health: Three lifecourse hypotheses.

Lindström, Martin; Fridh, Maria; Rosvall, Maria

Published in:Psychiatry Research

DOI:10.1016/j.psychres.2013.11.018

2014

Link to publication

Citation for published version (APA):Lindström, M., Fridh, M., & Rosvall, M. (2014). Economic stress in childhood and adulthood, and poorpsychological health: Three life course hypotheses. Psychiatry Research, 215(2), 386-393.https://doi.org/10.1016/j.psychres.2013.11.018

Total number of authors:3

General rightsUnless other specific re-use rights are stated the following general rights apply:Copyright and moral rights for the publications made accessible in the public portal are retained by the authorsand/or other copyright owners and it is a condition of accessing publications that users recognise and abide by thelegal requirements associated with these rights. • Users may download and print one copy of any publication from the public portal for the purpose of private studyor research. • You may not further distribute the material or use it for any profit-making activity or commercial gain • You may freely distribute the URL identifying the publication in the public portal

Read more about Creative commons licenses: https://creativecommons.org/licenses/Take down policyIf you believe that this document breaches copyright please contact us providing details, and we will removeaccess to the work immediately and investigate your claim.

1

Economic stress in childhood and

adulthood, and poor psychological

health: three life course hypotheses

Martin Lindströma,b,*, Maria Fridha, Maria Rosvalla,b

a Department of Clinical Sciences in Malmö

Lund University

S-205 02 Malmö

Sweden

b Centre for Economic Demography

Lund University

*Corresponding author

Word count (Main text): 4,497 Word count (Abstract): 172

2

Abstract

Investigations of mental health in a life course perspective are scarce. The aim is to

investigate associations between economic stress in childhood and adulthood, and

poor psychological health in adulthood with reference to the accumulation, critical

period and social mobility hypotheses in life course epidemiology. The 2008 public

health survey in Skåne is a cross-sectional postal questionnaire study. A random

sample was invited which yielded 28,198 respondents aged 18–80 (55%

participation). Psychological health was assessed with the GHQ12 instrument.

Logistic regression models were used to investigate the associations adjusting for age,

country of birth, socioeconomic status, emotional support, instrumental support and

trust, and stratifying by sex. The accumulation hypothesis was confirmed because

combined childhood and adulthood exposures to economic stress were associated with

poor psychological health in a graded manner. The social mobility hypothesis was

also confirmed. The critical period hypothesis was not confirmed because both

childhood and adulthood economic stress remained significantly associated with poor

psychological health in adulthood. Economic stress in childhood is associated with

mental health in adulthood.

Key words: Economic stress, mental health, GHQ12, life course perspective, social

capital, Sweden

3

1.Introduction

Poor psychological health is one of the chronic health problems with highest

prevalence globally. In Sweden, 39.8% of all newly granted sickness benefits among

men and 41.5% among women in 2006 were due to psychiatric disorders and

syndromes, including “burnout” syndrome (Danielsson, 2009). A 10-15% proportion

of all adults in Sweden are estimated to have mental health problems which would

motivate seeking help within the health care system (Persson, 2005). Poor

psychological health is not only a highly prevalent cause of chronic disease in the

general population. It is also an important explanation behind socioeconomic

differences in health in the general population. Socioeconomic differences in

psychiatric disorders have been observed internationally for decades (Stanfield and

Marmot, 1992). A recent study in southern Sweden has indicated that the non-manual

employees in higher positions category has a similar prevalence of poor psychological

health as non-manual employees in middle positions, non-manual employees in lower

positions, skilled manual workers, unskilled manual workers and self-employed (no

significant differences when analyzed with non-manuals in higher positions as

references group) among both men and women. In contrast, the early retired, the

unemployed, students and persons on long term sick leave have significantly higher

odds ratios of poor psychological health among both men and women (Lindström et

al., 2012). Poor psychological health in adulthood is also associated with other current

conditions in adulthood such as age, sex, country of birth (Lindström, 2004),

emotional support, instrumental support and generalized trust in other people

(Lindström et al., 2012). Trust in others may be regarded alternatively as a

psychological trait or as an aspect of social capital (Putnam, 2000).

Psychological health and socioeconomic differences in psychological health in

adulthood may, however, also be causally affected not only by current social and

economic conditions in adulthood but also by a wide variety of comparatively

prevalent childhood conditions. Childhood seems to be a particularly sensitive period

with regard to environmental disturbances which increase risk of depression (Heim et

al., 2010). Experiences of childhood adversity with psychological consequences

reaching into adulthood in the form of depression and anxiety are prevalent in western

countries. In the USA, the rate of child abuse and neglect reached 10.6% in 2007 (US

4

Department of Health and Human Services, 2009). It has been estimated that

approximately 30-40% of the risk of depression across the life course is genetically

determined, while the rest of the risk can be attributed to environmental factors

(Merkangas and Swendsen, 1997; Heim and Binder, 2012). The role of early life

environmental stress factors in developing major depression, apart from genetic

factors, has been demonstrated in twin studies (Kendler et al., 2000). Such childhood

stressors include physical, sexual, emotional and verbal abuse, neglect, social

deprivation, household dysfunctions including violence and witnessing violence,

poverty, parental separation, parental death or illness, disaster, substance abuse and

criminal activity (Brown et al., 2009). Recent studies suggest a substantially increased

risk of internalizing depressive and anxiety disorders into adulthood following early

life stress in childhood (Kendler et al., 2003; Nugent et al., 2011). Childhood sexual

or physical abuse was shown to be associated with increased risk of symptoms of

depression and anxiety, addiction, psychiatric admissions and suicide attempts

(McCauley et al., 1997). Poor paternal relationship or maternal overprotection is also

associated with increased risk of depression (Lizardi et al., 1995). In fact, there seems

to be a dose-response relationship between the severity of the experience of childhood

adversities and the severity of depressive episodes and overall lifetime experience of

chronic depression (Chapman et al., 2004) as well as between childhood adversities

and adult experience of mental health problems in general (Edwards et al., 2003).

In the latter two decades there has been a surge in life course research, i.e. research

concerning the influence of risk factors in early life on health later in life. The notion

that exposure to risk factors in early life in utero or in childhood may causally affect

health later in life was first empirically investigated by Barker in relation to the

metabolic syndrome, type II diabetes and cardiovascular diseases. Barker suggested

that the last trimester of life in utero was a “critical period” which if exposed to

growth retardation would eventually result in a number of increased risk factors and

diseases related to the metabolic syndrome (Barker, 1995; Barker, 1998). The critical

period hypothesis has later been empirically investigated with regard to a variety of

diseases other than the metabolic syndrome and its clinical consequences (Sahade et

al., 2011). The literature concerning childhood experiences of psychological and

psychosocial adversities and their effects on depression, anxiety and other mental

disorders in adulthood suggests that childhood may be regarded as a “critical period”

5

also for these conditions. Since Barker’s original work, an entire theoretical

framework, several models, a range of concepts connected with this theoretical

framework, and other complementary as well as competing life hypotheses have

emerged. Such hypotheses include the accumulation of risk hypothesis in several

forms including accumulation of risk over time and accumulation of risk over time by

clustering of risk factors over time, birth cohort effects, chains of risk models and

several variants of the critical period hypothesis (Ben-Schlomo and Kuh, 2002; Kuh

et al., 2003). Two main life course hypotheses other than the critical period hypothesis

are the accumulation of risk and social mobility hypotheses. The accumulation of risk

hypothesis suggests that exposures accumulate during the life course in order to

cumulatively increase the risk of chronic disease in adulthood (Hallqvist et al., 2004).

The social mobility hypothesis is more directly focused on life course social

epidemiology. The idea behind the social mobility hypothesis is that intra- and inter-

generational social mobility, mostly defined in terms of socioeconomic status (SES)

by occupational status, education, income or economic stress, will have an effect on

health in adulthood. This hypothesis also implies that social mobility per se should be

considered a potentially important social cause of disease (Lynch et al., 1994).

In this study the aim is to empirically test the critical period, accumulation and social

mobility hypotheses in the 2008 public health survey distributed to a random sample

of 18-80 year old adults in Skåne, southern Sweden, in relation to poor psychological

health in adulthood. A public health survey item concerning recalled economic stress

in childhood will be combined with an item concerning the current experience of

economic stress in adulthood. Previous studies have suggested the importance of

childhood poverty, social deprivation and various household dysfunctions as early life

stressors for mental health in childhood and adulthood (e.g. Brown et al., 2009), but

no study has empirically tested the association between both childhood and adulthood

economic stress and poor psychological health in relation to the accumulation, critical

period and social mobility hypotheses. The item concerning economic stress in

adulthood has been analyzed in relation to health and health related behaviours in

earlier studies based on earlier public health surveys in Sweden (Fritzell and

Burström, 2006). A previous study based on the 2008 Skåne survey has investigated

the associations between economic stress in childhood, economic stress in adulthood

and self-rated global health (Lindström, Hansen and Rosvall, 2012). A second study

6

has analyzed associations between economic stress in childhood and adulthood and

tobacco smoking (Lindström et al., 2013). We investigate whether economic stress in

childhood or economic stress in adulthood are critically associated with poor

psychological health in adulthood when investigated in the same model (critical

period), whether economic stress in childhood and economic stress in adulthood have

a graded association with poor psychological health in adulthood (accumulation), and

if social mobility is associated with poor psychological health in adulthood (social

mobility).

The aim is to investigate the associations between economic stress during childhood

and adulthood, and poor psychological health with reference to the accumulation,

critical period and social mobility hypotheses, including demographic,

socioeconomic, psychosocial factors and trust in the multiple adjusted models.

2. Methods and materials

2.1 Study population

The 2008 public health survey in Skåne in the southernmost part of Sweden is cross

sectional. A total of 28,198 persons randomly selected from the official population

registers of people living in Skåne born between 1928 and 1990 answered a postal

questionnaire in August-September 2008, and the participation rate was 55%. Two

reminder letters were sent. The study has been approved by the Ethical Committee at

Lund University, Sweden (No. 2010/343).

2.2 Definitions

2.2.1 Dependent variable

Self reported psychological health (GHQ12) includes twelve items reflecting different

aspects of psychological health. The items included in the GHQ12 are ”Have You

been able to concentrate on what You have been doing during the past weeks?”,

”Have You had problems with Your sleep during the past weeks?”, ”Do You feel that

You have been useful during the past weeks?”, “Have You been able to make

7

decisions in different areas during the past weeks?”, ”Have You felt tense during the

past weeks?”, ”Have You during the past weeks been able to appreciate what You

have been doing during the days?”, ”Have You been able to deal with Your problems

during the past weeks?”, ”Generally speaking, have You felt happy during the past

weeks?”. These eight items had four alternative answers: ”More/better than usually”,

”As usual”, ”Less than usual” and ”Much less than usual”. The items were

dichotomised with two alternatives denoting ”good” psychological health and two

alternatives denoting ”bad” psychological health. Four other items had somewhat

different alternative answers: ”Have You felt unable to deal with Your own personal

problems during the past weeks?”, ”Have You felt unhappy and depressed during the

past weeks?”, ”Have You lost faith in Yourself during the past weeks?” and ”Have

You felt worthless during the past weeks?”. The four alternative answers to these four

items were: ”Not at all”, ”No more than usually”, ”More than usually” and ”Much

more than usually”. The answers to these items were also dichotomised to denote

”bad” psychological health or ”good” psychological health. If three or more of all the

twelve items denoted ”bad” psychological health, general psychological health

(GHQ12) was denoted as ”bad”. This instrument for the measurement of

psychological health is the shortest (other GHQ measures contain for instance 28 or

60 items) but has still been shown to be a very robust measure of psychological health

(Goldberg et al., 1997).

2.2.2 Independent variables

Age was divided into 18-24, 25-34, 35-44, 45-54, 55-64 and 65-80 year age intervals.

All analyses were stratified by sex.

Born in Sweden/born in other country than Sweden. All participants born outside

Sweden were aggregated into a single category which was compared with the

category born in Sweden.

Socioeconomic status (SES) by occupation included the employed categories higher

non-manual employees, medium level non-manual employees, low level non-manual

employees, skilled manual works and unskilled manual workers, and also self-

8

employed/farmers. The substantial proportion outside the workforce entails early

retired before age 65 (health reasons or early retirement entitlement in the

employment contract reasons), on long term sick leave, unemployed, students, old age

pensioners above age 65, and unclassified.

Emotional support was assessed with the question “Do you feel that you have one or

some persons that can give you accurate personal support in order to cope with the

stress and problems of life”. It has four alternative answers: “Yes, I am absolutely

certain to get such support”, “Yes, possibly”, “Not certain”, and “No”. The three latter

alternatives were depicted as low emotional support.

Instrumental support stems from the question “Can you get help from one or some

persons in case of illness or practical problems (borrow minor things, help with

reparation, help to write a letter, get advice and information)?” It had the same

alternative answers as emotional support, and was dichotomized correspondingly.

Generalized trust in other people is a variable which assesses the individual’s

perception of generalized trust in other people (including unknown). It was appraised

by the item “Generally, you can trust other people” with the four alternative answers:

“Do not agree at all”, “Do not agree”, “Agree”, and “Completely agree”. The options

were dichotomized, the two first alternatives indicating low trust and the two latter

high.

Economic stress in childhood was appraised with the question “Did your family

experience economic hardship during your childhood?” with the three alternatives

“No, no significant problems” (category 1), “Yes, less severe problems and/or

problems during short time periods” (category 2) and “Yes, severe problems and/or

problems during long time periods” (category 3).

Economic stress in adulthood was assessed with the question “How often during the

past twelve months have you had problems paying your bills?” with the four

alternative answers “never” (category 1), “occasionally” (category 2), “every second

month” (category 3) and “every month” (category 3). The two latter options “every

9

second month” and “every month” were collapsed (into category 3) which yielded

three categories.

Economic stress in childhood and economic stress in adulthood (current situation)

were analyzed combined to address the three hypotheses concerning accumulation,

critical period and social mobility (Lindström, Hansen and Rosvall, 2012; Lindström

et al. 2013):

The accumulation hypothesis was investigated by adding the exposure to economic

stress in childhood and adulthood: respondents with no problems in childhood as well

as in adulthood being the most optimal combination (1+1), respondents with no

problems in either adolescence or adulthood combined with lesser (medium) problems

in either childhood or adulthood being the second best combination (1+2 or 2+1), the

least optimal combination being severe economic stress in both childhood and

adulthood (3+3). The (1+3), (3+1), (2+2) combinations were analyzed collapsed as

well as the (2+3) and (3+2) combinations, yielding a total five combinations.

The critical period hypothesis was tested by including both economic stress in

childhood and adulthood as two separate and categorized variables in the same

multiple models.

The social mobility hypothesis was investigated by analyzing the mobility from no

economic problems in childhood to either no problems (1+1), less frequent problems

(1+2) or severe problems in adulthood (1+3). The baseline economic stress in

childhood among respondents with less severe problems and/or problems during short

time periods (2+1, 2+2, 2+3) as well as with severe problems and/or problems during

long time periods (3+1, 3+2, 3+3) were analyzed similarly with economic stress in

childhood as baseline.

2.3 Statistics

Prevalence (%) of poor psychological health, age, country of birth, socioeconomic

status, emotional support, instrumental support, trust, economic stress in childhood

and economic stress in adulthood stratified by sex were calculated (table 1).

10

Prevalences (%) and odds ratios with 95% confidence intervals (OR:s, 95% CI) of

poor psychological health were calculated according to age, country of birth,

socioeconomic status, emotional support, instrumental support, trust, economic stress

in childhood and economic stress in adulthood (table 2). Crude, age-adjusted and

multiple adjusted odds ratios and 95% confidence intervals of poor psychological

health according to the accumulation hypothesis were calculated (table 3). Crude, age-

adjusted and multiple adjusted odds ratios and 95% confidence intervals of poor

psychological health according to the critical period hypothesis were calculated (table

4). Crude, age-adjusted and multiple adjusted odds ratios and 95% confidence

intervals of poor psychological health according to the social mobility hypothesis

were calculated (table 5). All statistical analyses in tables 2-5 were conducted in

logistic regression models and stratified by sex. The statistical analyses were

performed using the PASW software package version 20.0 (Norusis, 2012).

3. Results

Table 1 shows that 13.8% of the men and 18.2% of the women reported poor

psychological health. The prevalence of demographic, socioeconomic status,

emotional support, instrumental support, trust, and economic stress in childhood and

adulthood variables among men and women are also shown in table 1.

Table 2 shows that poor psychological health was more common among the young,

among those born abroad, with low socioeconomic status, unemployed, sick leave

pensioners, low emotional support, low instrumental support, low trust, economic

stress in childhood and economic stress in adulthood.

Table 3 shows that the crude odds ratios of poor psychological health were 1.50 (1.33-

1.69) in the (1+2, 2+1) accumulation group, 2.82 (2.48-3.20) in the (1+3, 3+1, 2+2)

accumulation group, 5.56 (4.61-8.70) in the (2+3, 3+2) accumulation group and 6.50

(4.83-8.73) in the (3+3) accumulation group compared to the (1+1) no life-course

economic stress accumulation reference group among men, and the crude odds ratios

were 1.41 (1.27-1.57), 2.58 (2.30-2.89), 3.11 (2.61.3.70), and 5.89 (4.47-7.77) among

women, respectively. These patterns remained across the multiple analyses (table 3).

11

Table 4 shows that the crude and age adjusted odds ratios of poor psychological

health according to economic stress in both childhood and adulthood were significant

for both men and women compared to the no stress alternatives, respectively, when

included in the same logistic regression model. The odds ratios were higher (higher

effect measure) for economic stress in adulthood than for economic stress in

childhood among both men and women. In the multiple adjusted model the odds

ratios remained significant compared to the no economic stress in childhood and

adulthood alternatives for men. In contrast, the less severe and/ or shorter period of

economic stress in childhood (category 2) became not significant, odds ratio 0.98

(0.88-1.09), among women. The severe and/or longer period of economic stress in

childhood (category 3) remained significant among women in the multiple adjusted

model, odds ratio 1.43 (1.24-1.66), as well as the odds ratios of poor psychological

health in the economic stress in adulthood categories (2 and 3) compared to the no

economic stress in adulthood reference category (category 1) among women.

Table 5 shows that social mobility by moving from no economic stress in childhood

to moderate economic stress in adulthood (category 1 to category 2) resulted in an

odds ratio 1.29 (1.06-1.57) of poor psychological health among men and 1.59 (1.36-

1.86) among women compared to the no economic stress in either childhood or

adulthood reference group (category 1 to category 1), respectively, while moving

from no economic stress in childhood to severe economic stress in adulthood

(category 1 to 3) resulted in an odds ratio 3.35 (2.64-4.25) of poor psychological

health among men and 3.13 (2.56-3.83) in the multiple models. Social mobility in the

form of moving from moderate economic stress in childhood to no economic stress in

adulthood (from category 2 to category 1) resulted in odds ratios 0.59 (0.47-0.73)

among men and 0.66 (0.54-0.81) among women compared to the moderate-moderate

(category 2 to category 2) reference group, while moving in the other direction from

moderate to severe economic stress (from category 2 to category 3) resulted in an

odds ratio 2.11 (1.58-2.83) among men and 1.16 (0.88-1.54) among women compared

to the moderate-moderate (from category 2 to category 2) reference group in the

multiple adjusted models. Compared to the age adjusted model, the odds ratio of poor

psychological health for women thus became not significant in the multiple adjusted

model. Finally, social mobility from severe to moderate economic stress (category 3

to category 2) resulted in odds ratios 0.64 (0.41-0.99) among men and 0.50 (0.33-

12

0.76) among women of poor psychological health compared to the severe-severe

(from category 3 to category 3) reference group, while social mobility from severe to

no economic stress (from category 3 to category 1) resulted in odds ratios 0.44 (0.30-

0.64) among men and 0.48 (0.34-0.68) among women of poor psychological health

compared to the severe-severe (from category 3 to category 3) reference group.

4. Discussion

This is the first study to investigate the accumulation, critical period and social

mobility life course hypotheses with regard to the association between economic

stress in childhood and adulthood and poor psychological health. The accumulation

hypothesis was fully confirmed regarding the relation between economic stress in

childhood and adulthood, and poor psychological health. The social mobility

hypothesis was also confirmed. In contrast, the fact that both economic stress in

childhood and economic stress in adulthood are significantly associated with poor

psychological health throughout the multiple logistic regression analyses indicates

that both periods are important, i.e. there seems to be no specific critical period for

economic stress and poor psychological adult health.

The three life course hypotheses are interconnected. In fact, the social mobility

hypothesis partly entails parts of the accumulation hypothesis, because when you, for

example, move down the social hierarchy, you add an exposure to low socioeconomic

status to your lifetime exposure and vice versa. The fact that we only have two

observation points in time, one retrospective and one current in the cross-section,

makes it harder to separate the test of the accumulation hypothesis from the test of the

social mobility hypothesis in the analyses (Singer and Willett, 2003). This also

necessitates the separation of measures of accumulation from measures of social

mobility, which we have done to the extent possible in this study. The critical period

hypothesis was not confirmed in this study, because both childhood and adulthood

economic stress were significantly and consistently associated with poor

psychological health when entered and analyzed simultaneously in the same crude,

age adjusted and multiple adjusted logistic regression models. This result seems to be

in conformity with previous literature which suggests that the importance of adverse

childhood conditions for adult psychological health problems does not rule out the

13

complementary importance of current social and economic circumstances and other

adversities in adulthood (Kendler et al., 2003; Nugent et al., 2011). In fact, the results

suggest that both childhood and adulthood are sensitive periods for the association

between economic stress and poor psychological health, because in contrast to the

concept “critical period” the concept “sensitive period” does not exclude the

possibility of two or more sensitive periods which each entail heightened sensitivity in

terms of increased importance of the association between exposure (economic stress)

and disease (poor psychological health) (Ben-Schlomo and Kuh, 2002).

The investigation of the effects of life course social and economic conditions on

health is sometimes stated to require three points of observation in time (Hallqvist et

al., 2004). In this study we have two points of observation in time. However, the

second and third points of observation in time are both in adulthood in the Hallqvist et

al. study (Hallqvist et al., 2004). In the present study we analyze, however, the

answers from adults in the broad age interval 18-80 years. When we stratify for age

by separately analyzing more narrow age intervals, we find the same results indicating

significant associations between both childhood and adulthood economic stress and

poor psychological health across all adult age intervals (not shown in tables). This

result suggests that there seems to be no particular “critical period” of economic stress

in adulthood for poor psychological health. Still, it should be noted that one previous

Swedish study which highlighted adolescence, not specifically analyzed in our study,

as one observation point in time found that the association between family economic

stress (cash margin) and adolescents’ health complaints largely disappeared when

adolescents’ own economic resources were controlled for (Aberg Yngwe and Östberg,

2013). Both economic stress in childhood and adulthood are significantly associated

with current poor psychological health regardless of adult age interval, but the fact

that the adolescence period is not included in the data and the fact that the study lacks

observation points in time precludes further interpretation.

The fact that both childhood and adulthood economic stress are significantly

associated with current poor psychological health in adulthood suggests that both

childhood and adulthood socioeconomic and psychosocial circumstances are

important to consider when health policy to promote good psychological health in

adulthood is discussed, elaborated and implemented. This also suggests that items

14

concerning economic stress in both childhood and adulthood should be included in

surveys which fully or partly concern adult psychological health. The result also

further underpins the importance for health and health policy of reducing economic

stress in childhood. Economic stress in childhood seems to be a significant risk factor

not only for poor psychological health in childhood but also in adulthood regardless

of adult age.

Logistic regression models and proportional hazards models are the most commonly

used regression models in psychiatric epidemiology. Still, these statistical models

have important limitations, because they e.g. neglect important information

concerning inter-individual variability (Bollen and Curran, 2006). Different analytical

tools which may broadly be labeled Structural Equation Modeling (SEM) are

therefore recommended in order to deal with these methodological limitations

(Papachristou et al., 2013). Such models utilizing more than two observation points

over time would also permit reliable application of more advanced analytical tools,

e.g. linear growth curve models which would determine the exact longitudinal

trajectory of economic stress, the rate of change (slope) and the level (intercept) of

economic stress across assessment points, the associations between the intercept and

the slope with the outcome variable, and the effects of additional covariates on the

risk trajectory in order to assess the life course hypotheses more accurately. Further

studies including longitudinal studies with three observation points in time or more

and utilizing Structural Equation Modeling are therefore warranted.

Causality cannot be inferred from cross-sectional studies. Still, one of the exposure

variables in this study is a retrospective self-reported item concerning economic stress

in childhood. Generally, the issue of causal inference is best addressed by

Randomized Control Triats (RCT:s). Still, in life course epidemiology an even more

important issue than the issue of temporality and longitudinal study design has been

the implicit assumption of no unmeasured confounding rather than the type of data

collection (Gilman, 2007). Furthermore, modern methods have been suggested to deal

with linear and non-linear Structural Equation Models (SEM:s) in order to make

causal inferences in the presence of unmeasured confounders (Pearl, 2000; Hernan

and Robins, 2006; Martens et al., 2006), a fact which even further warrants future

studies utilizing Structural Equation Modeling (SEM). Future studies may add more

15

longitudinal information by linking to register data, particularly concerning the

critical period hypothesis.

This study has focused on the effects of prolonged exposure when testing the

accumulation hypothesis and neglected accumulation by clustering, e.g. the

accumulation of social and psychosocial risk factors such as low emotional support,

low instrumental support and low trust on health. While such a cluster approach to

accumulation is outside the scope of our study, we acknowledge this alternative

approach to accumulation. The accumulation could for example also have been tested

regarding accumulation as a phenomenon of increasing burden of economic stress

during the life course, i.e. with one stable exposure category including (1+1), (2+2)

and (3+3) in terms of economic stress, one increasing exposure category including

(1+2), (1+3) and 2+3), and one decreasing exposure category (2+1), (3+2) and (3+1).

However, this option would imply a higher risk of mixing different associations and

effects, e.g. by including the (1+1) and (3+3) groups in the same category, and it

would probably also result in an even lower level of differentiation between the

accumulation and social mobility hypotheses.

The distribution of demographic, socioeconomic and social characteristics in a

previous public health survey with a similar response rate conducted in 2000 accorded

acceptably well with the distribution of these characteristics in the population of

Skåne in 2000 when compared with register statistics depicting the same

characteristics (Carlsson et al., 2006). Internationally unpublished data show some

under-representation in the age interval 18-34 years (22.0% among respondents but

29.0% in the original sample) and some corresponding over-representation in the 65-

80 year age interval (22.9% among respondents but 18.0% in the sample). Some

under-representation of men (45.1% among respondents and 50.0% in the sample)

and persons with low formal education (25.2% among respondents and 29.3% in the

sample) was also observed. The most serious under-representation (also present in the

data in the public health survey conducted in 2000) concerns people born outside

Europe (4.1% among respondents but 6.9% in the sample), a fact which would

introduce a problem in studies with the research focus on this particular group.

Comparisons for the 2008 study have given similar unpublished results. The risk of

selection bias thus seems to be limited although the participation rate was 55%.

16

Confounding by age, country of origin, socioeconomic status (by occupation and

employment status), social (emotional and instrumental) support and trust was

controlled for by adjustment in the age- and multiple analyses, and by stratifying for

sex.

The number of internally missing is comparatively small. In the analyses in tables 3-5

we have included all respondents which suggest that fewer respondents are included

in the final multiple than in the initial crude analyses because fewer respondents have

answered all the items/variables. In order to control the results of the crude and age-

adjusted analyses we restricted the number of respondents included to only those with

full information on all items/variables included in the final model. The results of these

alternative analyses suggest that missing data can be assumed to be missing under the

Missing At Random (MAR) or even the Missing Completely At Random (MCAR)

assumptions.

The GHQ12 instrument which includes twelve items for the measurement of

psychological health is the shortest of the General Health Questionnaire instruments

(other GHQ measures contain for instance 28 or 60 items) but still a valid and reliable

measure of psychological health (Goldberg et al, 1997). Some prior studies have

analyzed subjective economic hardships utilizing the same item or similar items and

shown significant associations with differing health outcomes (Fritzell and Burström,

2006). Information concerning economic stress in childhood is scarce, so the presence

of this variable in the data is a clear strength.

Conclusions: The accumulation and social mobility hypotheses were confirmed with

regard to poor psychological health. In contrast, the fact that both economic stress in

childhood and economic stress in adulthood are significantly associated with poor

psychological health indicates that both periods are important, i.e. there seems to be

no specific critical period.

17

Acknowledgements

This study was supported by the Swedish Research Council (Vetenskapsrådet)

(VR2014-2018-2674), Swedish Research Council Linnaeus Centre for Economic

Demography (VR 79), Swedish ALF Government Grant Dnr M M 2011/1816, and the

Research Funds of Southern University Hospital in Sweden.

18

References

Aberg Yngwe, M., Östberg, M., 2013. The family’s economic resources and

adolescents’ health complaints- do adolescents’ own economic resources matter?

European Journal of Public Health, 23(1), 24-29.

Barker, D.J.P., 1995. Fetal origins of coronary heart disease. British Medical Journal,

311, 171-174.

Barker, D.J.P., 1998. Mothers, babies and health in later life. Churchill Livingstone,

Edinburgh.

Ben-Schlomo, Y., Kuh, D., 2002. A life course approach to chronic disease

epidemiology: conseptual models, empirical challenges and interdisciplinary

perspectives. International Journal of Epidemiology, 31, 285-293.

Bollen, K.A., Curran, P.J., 2006. Latent Curve Models: A Stuctural Equation

Perspective. Wiley-Interscience, New Jersey.

Brown, D.W., Anda, R.F., Tiemeier, H., Felitti, V.J., Edwards, V.J., Croft, J.B. et al.,

2009 Adverse childhood experiences and the risk of premature mortality. American

Journal of Public Health, 37, 389-396.

Carlsson, F., Merlo, J., Lindström, M., Östergren, P.O., Lithman, T., 2006.

Representativity of a postal questionnaire survey in Sweden, with special reference to

ethnic differences in participation. Scandinavian Journal of Public Health, 34(2), 132-

139.

Chapman, D.I., Whitfield, C.L., Felitti, V.J., Dube, S.R., Edwards, V.J., Anda, R.F.,

2004. Adverse childhood experiences and the risk of depressive disorders in

adulthood. Journal of Affective Disorders, 82, 217-222.

Danielsson, M., 2009. Public Health Report 2009. The National Board on Health and

Welfare, Stockholm.

19

Edwards, V.J., Holden, G.W., Felitti, V.J., Anda, R.F., 2003. Relationship between

multiple forms of childhood maltreatment and adult mental health in community

respondents: results from the adverse childhood experiences study. American Journal

of Psychiatry, 160, 1453-1460.

Fritzell, S., Burström, B., 2006. Economic strain and self-rated health among lone and

couple mothers in Sweden during the 1990s compared to the 1980s. Health Policy, 79,

253.-264.

Gilman, S.E., 2007. Invited commentary: The life course epidemiology of depression.

American Journal of Epidemiology, 166, 1134-1137.

Goldberg, D.P., Gater, R., Sartorious, N., Ustun, T.B., Piccinelli, M., Gureje, O.,

Rutter, C., 1997. The validity of two versions of the GHQ in the WHO study of

mental illness in general health care. Psychology in Medicine, 27(1), 191-197.

Hallqvist, J., Lynch, J., Bartley, M., Lang, T., Blane, D., 2004. Can we disentangle

life course processes of accumulation, critical period and social mobility? An

Analysis of disadvantaged socio-economic positions and myocardial infarction in the

Stockholm Heart Epidemiology Program. Social Science and Medicine, 58, 1555-

1562.

Heim, C., Binder, E.B., 2012. Current research trends in early life vstress and

depression: review of human studies on sensitive periods, gene-environment

interactions, and epigenetics. Experimental Neurobiology, 233, 102-111.

Heim, C., Shugart, M., Craighead, W.E., Nemeroff, C.B., 2010. Neurobiological and

psychiatric consequences of child abuse and neglect. Developmental Psychobiology,

52, 671-690.

Hernan, M.A., Robins, J.M., 2006. Instruments for causal inference: An

epidemiologist’s dream? Epidemiology, 17, 360-372.

20

Kendler, K.S., Bulik, C.M., Silberg, J., Hettema, J.M., Myers, J., Prescott, C.A., 2000.

Childhood sexual abuse and adult psychiatric and substance use disorders in women:

an epidemiological and co-twin control analysis. Archives of General Psychiatry, 57,

953-959.

Kendler, K.S., Hettema, J.M., Butera, F., Gardner, C.O., Prescott, C.A., 2003. Life

event dimensions of loss, humiliation, entrapment, and danger in the prediction of

onsets of major depression and generalized anxiety. Archives of General Psychiatry,

60, 789-796.

Kuh, D., Ben-Schlomo, Y., Lynch, J., Hallqvist, J., Powe, C., 2003. Life course

epidemiology. Journal of Epidemiology and Community Health, 57, 778-783.

Lindström, M., 2004. Social capital, the miniaturization of community and self

reported global and psychological health. Social Science and Medicine, 59(3), 595-

607.

Lindström, M., Ali, S.M., Rosvall, M., 2012. Socioeconomic status, labour market

connection, and self-rated psychological health: The role of social capital and

economic stress. Scandinavian Journal of Public Health, 40(1), 51-60.

Lindström, M., Hansen, K., Rosvall, M., 2012. Economic stress in childhood and

adulthood, and self-rated health: a population based study concerning risk

accumulation, critical period and social mobility. BMC Public Health, 12, 761.

Lindström, M., Modén, B., Rosvall, M., 2013. A life course perspective on economic

stress and tobacco smoking: A population based study. Addiction, 108, 1305-1314.

Lizardi, H., Klein, D.N., Ouimette, P.C., Riso, L.P., Anderson, R.L., Donaldson, S.K.,

1995. Report of the childhood environment in early-onset dysthymia and episodic

major depression. Journal of Abnormal Psychology, 104, 132-139.

21

Lynch, J.W., Kaplan, G.A., Cohen, R.D., Kauhanen, J., Wilson, T.W., Smith, N.L.,

Salonen, J.T., 1994. Childhood and adult socioeconomic status as predictors of

mortality in Finland. Lancet, 343, 524-527.

Martens, E., Pestman, W., de Boer, A., Belitser, S.V., Klungel, O.H., 2006.

Instrumental variables: applications and limitations. Epidemiology, 17, 260-267.

McCauley, J., Kern, D.E., Kolodner, K., Dill, L., Schroeder, A.F., DeChant, H.K.,

Ryder, J., Derogatis, L.R., Bass, E.B., 1997. Clinical characteristics of women with a

history of childhood abuse: Unhealed wounds. Journal of the American Medical

Association, 277, 1362-1368.

Merkangas, K.R., Swendsen, J.D., 1997. Genetic epidemiology of psychiatric

disorders. Epidemiologic Reviews, 19, 144-155.

Norusis, M.J., 2012. PASW for Windows. Advanced Statistics. Release 20.0. SPSS,

Chicago.

Nugent, N.R., Tyrka, A.R., Carpenter, L.L., Price, L.H., 2011. Gene-environmental

interactions: early life stress and risk for depressive and anxiety disorders.

Psychopharmacology, 214, 175-196.

Papachristou, E., Frangou, S., Reichenberg, A., 2013. Expanding conceptual

frameworks: Life course risk modeling for mental disorders. Psychiatry Research 206,

140-145.

Pearl, J., 2000. Causality: Models, Reasoning, and Inference. New York: Cambridge

University Press.

Persson, G., 2005. Public Health 2005. The National Board on Health and Welfare,

Stockholm.

Putnam, R.D., 2000. Bowling Alone. The Collapse and Revival of American

Community. Simon & Schuster, New York.

22

Sahade, V., Franca, S., Badaro, R., Fernando Adan, L., 2011. Obesity and

postprandial lipemia in adolescents: Risk factors for cardiovascular diseases.

Endocrinologia y Nutricion, 59(2), 131-139.

Singer, J.D., Willett, J.B., 2003. Applied Longitudinal Data Analysis. Model Change

and Event Occurrence. Oxford University Press, Oxford, NewYork.

Stanfield, S.A., Marmot, M.G., 1992. Social class and minor psychiatric disorders in

British Civil Servants: a validated screening survey using the General Health

Questionnaire. Psychology in Medicine, 22, 739-749.

US Department of Health and Human Services, 2009. Administration on children,

youth and families. Child maltreatment 2007. US Government Printing Office,

Washington DC.

23

Table 1. Prevalence (%) of poor psychological health, age, country of birth, socioeconomic status, emotional support, instrumental support, generalized trust in other people, and economic stress in childhood and adulthood. Men (n = 12,726), women (n = 15,472), and total (n = 28,198). The public health survey in Skåne 2008.

Men (n = 12,726) Women (n = 15,472) Total (n = 28,198) Self reported psychological health (GHQ12)

Good 86.2 81.8 83.8 Poor 13.8 18.2 16.2 (Missing) (513) (627) (1140) Age 18-24 8.3 9.1 8.8 25-34 12.3 13.9 13.2 35-44 16.4 17.2 16.9 45-54 17.7 18.5 18.1 55-64 21.2 19.3 20.1 65-80 24.2 21.9 22.9 (Missing) (0) (0) (0) Country of birth Sweden 86.1 85.9 86.0 Other country 13.9 14.1 14.0 (Missing) (273) (282) (555) Socioeconomic status Higher non-manual 10.2 8.1 9.1 Medium non-manual 12.0 16.3 14.3 Lower non-manual 4.8 9.5 7.4 Skilled manual 10.7 8.7 9.6 Unskilled manual 11.6 11.1 11.3 Self-employed/farmer 7.7 3.7 5.5 Early retired 3.2 4.6 4.0 Unemployed 3.2 3.4 3.3 Student 4.9 6.7 5.9 Old age pensioner 26.2 23.2 24.6 Unclassified 4.7 3.4 4.0 Long term sick leave 0.9 1.3 1.1 (Missing) (212) (244) (456) Emotional support High 62.8 69.6 66.6 Low 37.2 30.4 33.4 (Missing) (289) (357) (646) Instrumental support High 71.3 76.6 74.2 Low 28.7 23.4 25.8 (Missing) (295) (338) (633) Trust High 66.1 64.3 65.2 Low 33.9 35.7 34.8 (Missing) (522) (685) (1207) Economic stress in childhood

No significant problem 63.2 62.5 62.8 Less severe and/or shorter period

27.1 27.7 27.4

Severe and/or longer period

9.7 9.7 9.7

(Missing) (341) (354) (695) Economic stress in adulthood

Never 79.5 76.5 77.8

24

Occasionally 14.1 15.7 15.0 Half the year 3.1 3.6 3.4 Every month 3.3 4.2 3.8 (Missing) (307) (335) (642)

25

Table 2. Prevalence (%) and odds ratios (OR, 95% CI) in bivariate analyses of poor psychological health (GHQ12) according to age, country of birth, socioeconomic status, emotional support, instrumental support, trust in other people (horizontal trust), and economic stress in childhood and adulthood. Men (n = 12,726) and women (n = 15,472). The public health survey in Skåne 2008. Men (n=12,726) Women (n=15,472) % OR(95%CI) % OR(95%CI) Age 18-24 19.5 1.00 31.5 1.00 25-34 18.4 0.93 (0.76-1.14) 23.8 0.68 (0.49-0.66) 35-44 15.3 0.74 (0.61-0.91) 20.8 0.57 (0.49-0.66) 45-54 16.2 0.80 (0.66-0.96) 16.8 0.49 (0.38-0.51) 55-64 12.4 0.58 (0.48-0.70) 15.2 0.39 (0.33-0.45) 65-80 8.1 0.36 (0.30-0.44) 10.8 0.26 (0.22-0.31) (Missing) (513) (627) Country of birth Sweden 12.5 1.00 16.8 1.00 Other country 22.9 2.08 (1.83-2.37) 26.3 1.77 (1.58-1.97) (Missing) (680) (804) Socioeconomic status Higher non-manual 13.1 1.00 16.7 1.00 Medium non-manual 11.7 0.88 (0.70-1.11) 15.5 0.91 (0.76-1.10) Lower non-manual 14.2 1.10 (0.83-1.46) 16.7 1.00 (0.81-1.23) Skilled manual 12.7 0.97 (0.77-1.22) 15.8 0.93 (0.71-1.15) Unskilled manual 11.7 0.88 (0.70-1.11) 18.8 1.15 (0.95-1.46) Self-employed/farmer 11.5 0.86 (0.66-1.11) 15.0 0.88 (0.67-1.17) Early retired 36.7 3.84 (2.95-5.02) 29.0 2.04 (1.63-2.55) Unemployed 37.2 3.93 (3.01-5.12) 40.9 3.44 (2.73-4.35) Student 20.7 1.73 (1.33-2.24) 27.8 1.92 (1.56-2.36) Old age pensioner 7.9 0.57 (0.46-0.70) 10.7 0.60 (0.49-0.72) Unclassified 17.1 1.37 (1.04-1.86) 22.5 1.44 (1.11-1.88) Long term sick leave 57.8 9.07 (5.99-13.72) 58.8 7.11 (5.16-9.79) (Missing) (649) (807) Emotional support High 9.7 1.00 13.7 1.00 Low 21.0 2.48 (2.24-2.76) 28.8 2.55 (2.34-2.78) (Missing) (681) (846) Instrumental support High 11.2 1.00 14.9 1.00 Low 20.7 2.07 (1.87-2.31) 29.2 2.35 (2.15-2.58) (Missing) (689) (826) Trust (horizontal) High 10.6 1.00 13.6 1.00 Low 19.9 2.10 (1.89-2.33) 25.7 2.26 (2.07-2.46) (Missing) (898) (1148) Economic stress in childhood No significant problem 11.4 1.00 16.0 1.00 Less severe and/or shorter period

16.2 1.50 (1.34-1.69) 20.1 1.32 (1.20-1.45)

Severe and/or longer period 23.2 2.35 (2.01-2.74) 26.2 1.86 (1.63-2.12) (Missing) (745) (871) Economic stress in adulthood Never 10.8 1.00 14.2 1.00 Occasionally 20.7 2.16 (1.89-2.47) 26.2 2.14 (1.93-2.38) Half the year 34.2 4.31 (3.44-5.39) 32.3 2.88 (2.38-3.48) Every month 40.1 5.53 (4.46-6.85) 44.8 4.88 (4.13-5.78) (Missing) (705) (844)

26

Table 3. Prevalence (%) and odds ratios (OR, 95% CI) in crude, age-adjusted and multiple adjusted analyses of poor psychological health according to economic stress risk accumulation (childhood+ adulthood combined). Men (n = 12,726) and women (n = 15,472). The public health survey in Skåne 2008. Men Risk accumulation

% OR(95% CI)a OR(95% CI)b OR(95% CI)c

Lowest (1+1) 7.1 1.00 1.00 1.00 (1+2) or (2+1) 13.5 1.50 (1.33-1.69) 1.51 (1.34-1.70) 1.33 (1.17-1.51) (1+3), (2+2) or (3+1)

22.2 2.82 (2.48-3.20) 2.85 (2.50-3.24) 2.31 (2.01-2.65)

(2+3) or (3+2) 35.6 5.56 (4.61-8.70) 5.41 (4.49-6.53) 4.06 (3.32-4.98) Highest (3+3) 42.9 6.50 (4.83-8.73) 6.40 (4.75-8.61) 4.43 (3.20-6.13) Women Risk accumulation

% OR(95% CI)a OR(95% CI)b OR(95% CI)c

Lowest (1+1) 13.0 1.00 1.00 1.00 (1+2) or (2+1) 18.1 1.41 (1.27-1.57) 1.40 (1.25-1.55) 1.24 (1.11-1.39) (1+3), (2+2) or (3+1)

27.7 2.58 (2.30-2.89) 2.47 (2.20-2.77) 1.96 (1.73-2.22)

(2+3) or (3+2) 32.3 3.11 (2.61-3.70) 2.89 (2.43-3.45) 1.85 (1.53-2.25) Highest (3+3) 46.8 5.89 (4.47-7.77) 5.61 (4.24-7.43) 3.42 (2.53-4.63) a Crude. b Adjusted for age. c Adjusted for age, country of birth, socioeconomic status, emotional support, instrumental support and trust..

27

Table 4. Odds ratios (OR, 95% CI) in crude, age-adjusted and multiple adjusted analyses of poor psychological health according to economic stress critical period (childhood+ adulthood included as separate variables in the same model). Men (n = 12,726) and women (n = 15,472). The public health survey in Skåne 2008. Men Critical period

OR(95% CI)a OR(95% CI)b OR(95% CI)c

Economic stress in childhood

No significant problem (1)

1.00 1.00 1.00

Less severe and/or shorter period (2)

1.59 (1.42-1.77) 1.64 (1.47-1.82) 1.50 (1.34-1.68)

Severe and/or longer period (3)

1.98 (1.71-2.30) 2.14 (1.84-2.48) 1.83 (1.56-2.14)

Economic stress in adulthood

Never (1) 1.00 1.00 1.00 Occasionally (2)

1.75 (1.55-1.98) 1.62 (1.42-1.83) 1.45 (1.27-1.66)

Half the year/ every month (3)

4.24 (3.66-4.90) 3.96 (3.42-4.59) 3.16 (2.70-3.70)

Women OR(95% CI)a OR(95% CI)b OR(95% CI)c Economic stress in childhood

No significant problem (1)

1.00 1.00 1.00

Less severe and/or shorter period (2)

1.10 (1.00-1.21) 1.14 (1.03-1.26) 0.98 (0.88-1.09)

Severe and/or longer period (3)

1.60 (1.40-1.83) 1.74 (1.52-2.00) 1.43 (1.24-1.66)

Economic stress in adulthood

Never (1) 1.00 1.00 1.00 Occasionally (2)

2.01 (1.81-2.24) 1.73 (1.55-1.93) 1.50 (1.33-1.68)

Half the year/ every month (3)

3.70 (3.25-4.22) 3.29 (2.88-3.76) 2.46 (2.13-2.84)

a Crude. b Adjusted for age. c Adjusted for age, country of birth, socioeconomic status, emotional support, instrumental support and trust.

28

Table 5. Prevalence (%) and odds ratios (OR, 95% CI) in crude, age-adjusted and multiple adjusted analyses of poor psychological health according to social mobility (childhood to adulthood). Men (n = 12,726) and women (n = 15,472). The public health survey in Skåne 2008. Men Social mobility (childhood-adulthood)

% OR(95% CI)a OR(95% CI)b OR(95% CI)c

No-Never (1 to 1)

9.4% 1.00 1.00 1.00

No-occasionally (1 to 2)

17.5% 1.58 (1.32-1.90) 1.46 (1.22-1.76) 1.29 (1.06-1.57)

No-half the year/every month (1 to 3)

32.6% 4.25 (3.41-5.29) 4.12 (3.31-5.14) 3.35 (2.64-4.25)

Social mobility (childhood-adulthood)

% OR(95% CI)a OR(95% CI)b OR(95% CI)c

Less severe-Never (2 to 1)

12.0% 0.49 (0.40-0.60) 0.54 (0.44-0.66) 0.59 (0.47-0.73)

Less severe-occasionally (2 to 2)

22.2% 1.00 1.00 1.00

Less severe-half the year/ every month (2 to 3)

39.7% 2.50 (1.90-3.28) 2.49 (1.89-3.27) 2.11 (1.58-2.83)

Social mobility (childhood-adulthood)

% OR(95% CI)a OR(95% CI)b OR(95% CI)c

Severe-Never ( 3 to 1)

17.9% 0.33 (0.23-0.45) 0.36 (0.26-0.50) 0.44 (0.30-0.64)

Severe-occasionally (3 to 2)

29.7% 0.53 (0.36-0.80) 0.55 (0.37-0.83) 0.64 (0.41-0.99)

Severe-half the year/ every month (3 to 3)

42.9% 1.00 1.00 1.00

Women Social mobility (childhood-adulthood)

% OR(95% CI)a OR(95% CI)b OR(95% CI)c

No-Never (1 to 1)

13.0% 1.00 1.00 1.00

No- 23.8% 2.02 (1.74-2.33) 1.76 (1.52-2.04) 1.59 (1.36-1.86)

29

occasionally (1 to 2) No-half the year/every month (1 to 3)

40.0% 4.47 (3.72-5.38) 4.05 (3.36-4.88) 3.13 (2.56-3.83)

Social mobility (childhood-adulthood)

% OR(95% CI)a OR(95% CI)b OR(95% CI)c

Less severe-Never (2 to 1)

15.6% 0.46 (0.39-0.56) 0.58 (0.48-0.70) 0.66 (0.54-0.81)

Less severe-occasionally (2 to 2)

28.6% 1.00 1.00 1.00

Less severe-half the year/ every month (2 to 3)

33.8% 1.36 (1.05-1.75) 1.45 (1.12-1.87) 1.16 (0.88-1.54)

Social mobility (childhood-adulthood)

% OR(95% CI)a OR(95% CI)b OR(95% CI)c

Severe-Never (3 to 1)

20.0% 0.30 (0.22-0.41) 0.33 (0.24-0.69) 0.48 (0.34-0.68)

Severe-occasionally (3 to 2)

30.3% 0.47 (0.33-0.69) 0.47 (0.33-0.69) 0.50 (0.33-0.76)

Severe-half the year/ every month (3 to 3)

46.8% 1.00 1.00 1.00

a Crude. b Adjusted for sex and age. c Adjusted for sex, age, country of birth, socioeconomic status, emotional support and, instrumental support and trust. 26,856 respondents included in analyses, 1342 respondents missing values.